MICROPROCESSOR INPUT/OUTPUT

4.1 Introduction
4.2 Simple I/O Devices
4.3 Programmed I/O
4.4 Unconditional and Conditional
Programmed I/O
4.5 Interrupt I/O
4.6 Direct Memory Access (DMA)
Summary of I/O

The technique of data transfer between a

microcomputer and an external device is
called input/output (I/O).
Peripherals are the I/O devices that
connected to a microcomputer and provide
an efficient means of communication
between the microcomputer and the
outside world.

Because the characteristics of I/O devices are

normally different from those of a
microcomputer like (speed and word length)
we need interface hardware circuitry between
the microcomputer and I/O devices
Interface hardware provide all input and
output transfers between the microcomputer
and peripherals by using an I/O bus.
An I/O bus carries three types of signals:
device address, data, and command.

There are three ways of transferring data

between a microcomputer and physical
I/O devices: programmed I/O, interrupt I/O
and direct memory access.

Programmed I/O, the microprocessor

executes a program to perform all data
transfers between the microcomputer and
the external device.
The main characteristic of this type of I/O
technique is that the external device carries
out the functions dictated by the program
inside the microcomputer memory.

Interrupt I/O, an external device can force

the microprocessor to stop executing
the current program temporarily so that it
can execute another program known as an
interrupt service routine.
After completing this program, a return
from interrupt instruction can be executed
at the end of the service routine to return
control at the right place in the main
program.

Direct memory access (DMA) is a type of I/O

technique in which data can be transferred
between microcomputer memory and an
external device such as the hard disk,
without microprocessor involvement.
A special chip called the DMA controller
chip is typically used with the
microprocessor for transferring data using
DMA.

For reading only

A microcomputer communicates with an

external device via one or more registers
called I/O ports using programmed I/O.
I/O ports are usually of two types.
1. For one type, each bit in the port can be
configured individually as either input or
output.
2. For the other type, all bits in the port can be
set up as all parallel input or parallel output
bits.

10

Command or data-direction register

Is a register used to configure each port as an
input or output port. The port contains the
actual input or output data. The data
direction register is an output register and
can be used to configure the bits in the port
as inputs or outputs.

11

As an example, if an 8-bit data-direction

register contains 34H (34 Hex), the
corresponding port is defined as shown
in Figure 4.4.

12

For parallel I/O, there is only one data

direction register for all ports. A particular
bit in the data direction register configures
all bits in the port as either inputs or
outputs.

13

I/O ports are addressed using either standard

I/O or memory-mapped I/O techniques.
1. Standard I/O or port I/O (called isolated I/O
by Intel) uses an output pin such as
the M / IO pin
2. In memory-mapped I/O, the microprocessor
does not use the M / IO control pin. Instead,
the microprocessor uses an unused address
pin to distinguish between memory and I/O.

14

15

There are typically two ways in which

programmed I/O can be utilized: unconditional
I/O and conditional I/O.
The microprocessor can send data to an
external device at any time using
unconditional I/O.
In conditional I/O, the microprocessor outputs
data to an external device via handshaking.
This means that data transfer occurs via the
exchange of control signals between the
microprocessor and an external device.
16

17

The concept of conditional I/0 will now be

demonstrated by means of data transfer
between a microprocessor and an analogto-digital (A/D) converter.

18

19

20

A disadvantage of conditional programmed

I/O is that the microcomputer needs to
check
the status bit (a conversion complete signal
of the A/D converter) by waiting in a loop.
Interrupt I/O is a device-initiated I/O
transfer. The external device is connected
to a pin called the interrupt (INT) pin on the
microprocessor chip.

21

How Interrupt work ??

1. When the device needs an I/O transfer with
the microcomputer, it activates the
interrupt pin of the processor chip.
2. The microcomputer usually completes the
current instruction and saves the contents
of the current program counter and the
status register in the stack.

22

How Interrupt work ??

3. The microcomputer then loads an address
automatically into the program counter to
branch to a subroutine-like program called
the interrupt service routine
4. The last instruction of the service routine
is a RETURN, which is typically similar in
concept to the RETURN instruction used at
the end of a subroutine.

23

Assume that the microcomputer is 68000

based and is executing the following
instruction sequence:

24

Assume that the address of service routine,

is $4000 and that the user writes a service
routine to input the A/D converters
output as follows:

25

There are typically three types of interrupts:

external interrupts, traps or internal
interrupts,and software interrupts
External interrupts can be divided further
into two types: maskable and nonmaskable.
Nonmaskable interrupt cannot be enabled
or disabled by instructions, whereas a
microprocessors instruction set contains
instructions to enable or disable maskable
interrupt

26

A nonmaskable interrupt is typically used as

a power failure interrupt.

27

Internal interrupts, or traps, are

activated internally by exceptional
conditions such as overflow, division by
zero, or execution of an illegal op-code.
Many microprocessors include software
interrupts, or system calls. When one of
these instructions is executed, the
microprocessor is interrupted and serviced
similarly to external or internal interrupts.

28

Software interrupt instructions are

normally used to call the operating system.
These instructions are shorter than
subroutine calls, and no calling program is
needed to know the operating systems
address in memory.

29

interrupt address vector

Is the technique used to find the starting
address of the service routine

30

When a microprocessor is interrupted, it

normally saves the program counter (PC)
and
the status register (SR) onto the stack so
that the microprocessor can return to the
main
program with the original values of PC and
SR after executing the service routine.

31

It is a special mechanism necessary to handle

interrupts from several devices that share
one of these interrupt lines.
There are two ways of servicing multiple
interrupts:
1. Polled technique
2. Daisy chain technique.

32

Polled Interrupts
The microprocessor responds to an interrupt
by executing one general service routine for
all devices.
The priorities of devices are determined by
the order in which the routine polls each
device.

33

Polled Interrupts

34

Polled Interrupts
several external devices (device 1, device 2,. . . ,
device N) are connected to a single interrupt line
of a microprocessor via an OR gate
When one or more devices activate the INT line
HIGH, the microprocessor pushes
the PC and SR onto the stack.
The user can write a program at this address to
poll each device, starting with the highestpriority device, to find the source of the interrupt.

35

Polled Interrupts
Suppose that the devices in Figure 4.10 are
A/D converters.

36

Polled Interrupts (example A/D converter)

Suppose that the user assigns device 2 the higher

priority.
1. When the Conversion complete signals from
device 1 and/or 2 become HIGH, the processor is
interrupted.
2. In response, the microprocessor pushes the PC and
SR onto the stack and loads the PC with the interrupt
address vector defined by the manufacturer.
3. If this device 2 has generated an interrupt, the
output (PB 1) of the AND gate in Figure 4.11
becomes HIGH

37

Polled Interrupts
Polled interrupts are slow, and for a large
number of devices the time required
to poll each device may exceed the time to
service the device. In such a case, a faster
mechanism, such as the daisy chain
approach, can be used.

38

Daisy Chain Interrupts

Devices are connected in daisy chain fashion, as

shown in Figure 4.12, to set up priority systems.
Suppose that one or more devices interrupt the
processor. In response, the micropro-cessor
pushes the PC and SR onto the stack and,
generates an interrupt acknowledge (INTA) signal
to the highest-priority- device (device 1 in this
case).

39

Daisy Chain Interrupts

If this device has generated the interrupt, it w ill

accept the INTA;
otherwise, it will pass the INTA onto the next
device until the INTA is accepted.
Once accepted, the device provides a means for
the processor to find the interrupt address vector
by using external hardware.

40

Daisy Chain Interrupts

41

Daisy Chain Interrupts (example A/D

Converter)

42

Daisy Chain Interrupts (example A/D Converter)

When the conversion complete signal goes HIGH,

the microprocessor is interrupted through the INT line.
2. The microprocessor pushes the PC and SR. It then
generates a LOW at the interrupt acknowledge (INTA)
for the highest-priority device. Device #1 has the
highest priority; it is the first device in the daisy chain
configuration to receive INTA .
3. If A/D converter 1 has generated the conversion
complete HIGH, the output of the AND gate in Figure
4.13 becomes HIGH.
1.

43

Direct memory access (DMA) is a technique that

transfers data between a microcomputers
memory and an I/O device without involving the
microprocessor.
The DMA technique uses a DMA controller chip for
the data transfer operations.

44

DMA controller chip Functions

1. The

I/O devices request DMA operation via the DMA

request line of the controller chip.
2. The controller chip activates the microprocessor HOLD
pin, requesting the microprocessor to release the bus.
3. The microprocessor sends HLDA (hold acknowledge)
back to the DMA controller, indicating that the bus is
disabled. The DMA controller places the current value
of its internal registers, on the system bus and
sends a DMA acknowledge to the peripheral device.

45

DMA controller chip Types

There are three basic types of DMA:

1. Block transfer DMA
2. Cycle stealing DMA
3. Interleaved DMA.

46

DMA controller chip Types

Block transfer DMA

The DMA controller chip takes over the bus from the
microcomputer to transfer data between the
microcomputer memory and the I/O device. The
microprocessor has no access to the bus until the
transfer is completed.

47

DMA controller chip Types

cycle stealing DMA

Data transfer between the microcomputer memory
and an I/O device occurs on a word-by-word basis

48

o
o

DMA controller chip Types

cycle stealing DMA
The microprocessor is generated by ANDing an INHIBIT
signal with the system clock.
The DMA controller controls the INHIBIT line.
During normal operation, the INHIBIT line is HIGH,
providing the microprocessor clock.
When DMA operation is desired, the controller makes the
INHIBIT line LOW for one clock cycle. The microprocessor
is then stopped completely for one cycle.
Data transfer between the memory and I/O takes place
during this cycle.

49

DMA controller chip Types

cycle stealing DMA

This method is called cycle stealing because the
DMA controller takes away or steals a cycle
without microprocessor recognition. Data transfer
takes place over a period of time.

50

DMA controller chip Types

interleaved DMA,
The DMA controller chip takes over the system bus
when the microprocessor is not using it.
The DMA controller chip identifies these cycles and
allows transfer of data between memory and the
I/O device. Data transfer for this method takes
place over a period of time.

51

DMA controller chip (How it works)

52

DMA controller chip (How it works)

The DMA controller chip usually has at least three registers normally
selected by the controller's register select (RS) line:
1. An address register,
2. A terminal count register,
3. And a status register.
Both the address and terminal counter registers are initialized by the
microprocessor. The address register contains the starting address of
the data to be
transferred, and the terminal counter register contains the block to be
transferred. The
status register contains information such as completion of DMA
transfer. Note that the
DMA controller implements logic associated with data transfer in
hardware to speed up the
DMA operation.

53

DMA controller chip (How it works)

The address register contains the starting

address of the data to be transferred.
The terminal counter register contains
the block to be transferred.
The status register contains information
such as completion of DMA transfer.
Note that the DMA controller implements
logic associated with data transfer in
hardware to speed up the DMA operation.
54

55